Nodes

A node is any computer in your WATCHOUT network that runs WATCHOUT software. A node can host any combination of services — Director, Asset Manager, Runner, Producer — and can change roles between shows. In a typical production the operator designs the show on a Producer node, and other nodes render through their Runner service, driving projectors, LED walls, monitors, audio systems, and capture devices.

What a Runner node does

When a node starts, its management service announces it on the network so Producer can discover it. Once the Director assigns work to the node (based on which displays are mapped to its name), the Runner service:

  1. Connects to the Director to receive playback instructions.
  2. Downloads required assets from the Asset Manager and caches them locally for instant playback.
  3. Renders content to its assigned display outputs using the node's GPU, and routes audio to its audio devices.
  4. Reports status back to Producer — download progress, playback health, GPU usage, drive space, and network throughput.

A single node can drive multiple outputs if it has multiple GPU outputs, SDI cards, ST 2110 channels, or NDI® streams configured.

Hardware considerations

Rendering nodes are usually dedicated computers built for media playback:

  • GPU — the main factor in rendering performance. Choose a GPU that handles the resolution, codec, and layer count the show needs. WATCHOUT runs on any GPU and uses hardware video decoding whenever the driver exposes it. NVIDIA GPUs add capabilities other vendors do not provide: hardware frame synchronization (Quadro Sync), GPUDirect transfers between the GPU and SDI / ST 2110 cards, WATCHOUT-managed EDID emulation, and driver-level color configuration.
  • Storage — use a fast SSD for the working directory that holds the node's assets. The high-fidelity, low-compression formats it uses demand a lot of disk bandwidth. Size the drive to the formats and resolutions your show plays — see the bandwidth and storage calculator.
  • Network — Gigabit Ethernet minimum, 10GbE for installations with many large assets or frequent show changes. NDI® video runs over the network too. A node can use several network adapters at once — Art-Net, Dante, and NDI are often kept on separate networks. See Network Requirements.
  • CPU — important for playback. It moves frame data to and from the GPU, and most video formats are partially decoded on the CPU. It also handles audio rendering and NDI processing.

For dedicated installations, Dataton offers WATCHPAX hardware — purpose-built nodes pre-configured with WATCHOUT software and a working directory set up.

Setup checklist

  1. Install WATCHOUT node software. The node management service configures the firewall rules automatically when it starts (see Firewall Configuration).
  2. Verify network connectivity. Put the node on the same subnet as Producer and the other nodes.
  3. Confirm the node appears in Producer. Open Window > Nodes and check it shows up with its name and services.
  4. Assign displays to the node. In the display's properties, set the Node field of each display to the node's name. This tells the Director which Runner renders each output.
  5. Assign an Asset Manager so the node can download media.
  6. Assign a Director so the node receives playback commands.
  7. Verify asset download in the Nodes window after the Director starts.
  8. Test playback. Run a timeline and confirm content appears on the physical outputs.

System readiness checklist

Before running a show, confirm the system is ready:

  1. All nodes visible. Set the Nodes window filter to Referred by Show and confirm every node the show expects is listed.
  2. Correct services running — Director, Asset Manager, and Runners active on the intended nodes.
  3. No version warning. A warning icon means a node runs a different WATCHOUT version than Producer. Match the versions — see Software Updates.
  4. No multi-show warning. A warning icon means a Runner is running a different show than the Director controlling it. This usually happens when more than one Director is active and their shows reference the same nodes. Make sure a single Director controls each node.
  5. Asset transfers complete. While an asset transfers, a placeholder renders in its place. Programming continues while transfers run. Before the show, let them finish so every asset is present.
  6. Display routing correct. Each display assigned to the right Runner.
  7. Clocks in sync. Each node's NTP offset should be near zero — see Time Synchronization.

Startup Action and node status

The node running the Director can load a show automatically on boot — useful for unattended installations that must start playback after a power cut. Set this with the Director's Startup Action. See Node Management and Maintenance.

In the Nodes window, watch each node's services, version, NTP offset, download progress, and resource use. See The Nodes Window for the full status-indicator list.

Cloning node disk images

Each WATCHOUT node needs a unique machine ID, which WATCHOUT generates the first time it runs. If you clone a disk image after WATCHOUT has already run, every clone inherits the same ID, and the nodes collide in the Nodes window — they may appear as duplicates or behave unpredictably.

Capture the disk image before WATCHOUT runs for the first time. Each node then generates its own ID on first launch. If an image was captured after WATCHOUT had already run, rebuild it from a clean installation.

Locking display output channel numbers

By default, WATCHOUT derives a display's channel number from the Windows display ID. Windows can renumber these IDs when a cable is unplugged and reconnected, when a display is power-cycled, or when the monitor topology changes. The new numbering then moves WATCHOUT content to the wrong physical output.

The reliable fix is EDID emulation: make Windows believe a monitor is permanently attached to each port. The desktop layout then never changes. Once Windows always sees the same displays, WATCHOUT's channel numbers stay stable.

  • NVIDIA outputs — WATCHOUT can emulate EDID directly from Device Properties. Capture each display's EDID once and apply the saved asset on the port. See GPU Output → EDID.
  • AMD outputs — enable EDID emulation in the AMD driver settings (Adrenalin / Radeon Pro). WATCHOUT does not manage EDID on AMD hardware.

Environment variable fallback

If EDID emulation is not practical, pin channel numbers to physical GPU connectors with a Windows environment variable on the node:

USE_GPU_CONNECTORS_AS_CHANNELS=TRUE

With this set, WATCHOUT numbers channels by the physical port each display is wired to instead of by Windows display ID. Each channel stays bound to the same connector regardless of how Windows reorders displays. The mapping covers NVIDIA outputs. Non-NVIDIA outputs are auto-assigned channels after the NVIDIA range.

WATCHPAX units ship with this variable set and a port mapping pre-configured by Dataton. On these units the mapping is authoritative — only mapped outputs receive WATCHOUT content. Non-mapped outputs (such as an Intel HDMI used for Producer access) are excluded automatically.

After setting or changing this variable, restart WATCHOUT for it to take effect.

Setting environment variables is more involved than EDID emulation. Prefer EDID, and use this variable only when EDID emulation cannot cover all outputs.

NDI® is a registered trademark of Vizrt NDI AB.

Previous Next